Notes on H. Naledi Analysis: Get the data

MeshLab worklflow

1. Open model
2. Click ‘pick points’ button
3. Right-click to add points, click ‘not editing’ button to rotate and move model
4. In the pick points ‘form’ window, click Save to save points into a pp file (XML)

Getting data from Google Drive

Our data are stored on Google Drive. We can access it for analysis using R

library(googledrive)

# here is the ID of our g-drive data folder
data_folder_on_googl_drv_url <- "1IipPeG0bQF92DttpSiCZyiv1H_95UENP"

# get the IDs for each pp file on g-drive
data_files_on_googl_drv <- 
  drive_ls(as_id(data_folder_on_googl_drv_url))

# download them to our local folder
pwalk(data_files_on_googl_drv,
    ~drive_download(as_id(..2), # refer to column 2 for the ID
                    # refer to column 1 for the name
                    path = here("analysis", "data", "raw-data", ..1), 
                    overwrite = TRUE))

Importing Meshlab data into R

Here’s the function for reading in pp files

# https://rgriff23.github.io/2017/05/07/read-meshlab-pickedpoints.html
# function to read Meshlab *.pp files into R
read.pp <- function (file) {
    file <- readLines(file)
    lines <- file[grep("point", file)]
    x <- strsplit(lines, "x=\"")
    y <- strsplit(lines, "y=\"")
    z <- strsplit(lines, "z=\"")
    name <- strsplit(lines, "name=\"")
    mat <- matrix(0, length(x), 3)
    r <- c()
    for (i in 1:length(lines)) {
        mat[i,1] <- as.numeric(strsplit(x[[i]][2], "\"")[[1]][1])
        mat[i,2] <- as.numeric(strsplit(y[[i]][2], "\"")[[1]][1])
        mat[i,3] <- as.numeric(strsplit(z[[i]][2], "\"")[[1]][1])
        r <- c(r, strsplit(name[[i]][2], "\"")[[1]][1])
    }
    
  df <- as_tibble(mat, .name_repair = make.names )
    df[ ,ncol(df) + 1] <- r 
    colnames(df) <- c( "x","y","z","landmark")
    df <- df[ , c(ncol(df), 1:ncol(df)-1) ]
    return(df)
}

Here we read in all the pp files that we downloaded from g-drive

# get list of the local data files
local_data_files <- 
  list.files(path = here("analysis", "data", "raw-data"),
             pattern = ".pp$",
             full.names = TRUE)

# read these into R, and assign names to each data frame
nested_df_of_pp_data <- 
  map(local_data_files, 
      read.pp) %>% 
  set_names(basename(local_data_files)) %>% 
  bind_rows(.id = 'specimen') %>% 
  mutate(landmark = as.numeric(landmark)) %>% 
  nest(-specimen)

# write an object we can reuse
saveRDS(nested_df_of_pp_data, 
        here("analysis", "data", "derived-data", 'nested_df_of_pp_data.rds'))

Take a quick look at one specimen:

library(plotly)
nested_df_of_pp_data %>% 
  filter(specimen == 'austro_afarensis_al-137-48a_21-landmarks.pp') %>% 
  unnest() %>% 
  plot_ly(.,
         x = ~x, 
         y = ~y, 
         z = ~z, 
         color = ~landmark) %>%
  add_markers() 

One point is out of place here, we may need to exclude it.